The present invention relates generally to testing techniques for electronic devices. More particularly, the present invention relates to a technique for measuring the frequency response of a transimpedance amplifier.
A transimpedance amplifier (TIA) operating on the output of a photodiode forms a fiber-optic receiver. In data applications, the output of the TIA is limited and subsequently processed by a clock and data recovery unit to obtain the original data. In such a system, the bandwidth of the TIA plays an important role. If the bandwidth is too narrow, the data will be corrupted by intersymbol interference. If the bandwidth is too wide, the TIA will generate excess noise that can also corrupt the data. From a communications theory viewpoint, the TIA bandwidth should be matched to that of the data channel.
For practical reasons, it is desirable to include the TIA and the limiter in a single integrated circuit (IC) device. For example, FIG. 1 is a schematic diagram of an IC package 100 having a TIA 102 and an integrated limiter 104. Such an integrated assembly is referred to herein as a TIA/limiter. The prior art contains a number of TIA/limiter designs, and the operation of TIA/limiters will not be described in detail herein. Briefly, TIA 102 generates an output that is proportional to the input current, and limiter 104 clips the TIA output signal at specific high and low voltage levels. The combined nature of an integrated TIA/limiter device makes it difficult (if not impossible) to perform direct measurements of the TIA itself (such as frequency response or bandwidth measurements).
One approach to testing the frequency response of a TIA implemented in an integrated TIA/limiter device is to include a TIA output test pin on the IC package. With such a configuration, conventional frequency response testing of the TIA is possible. However, an additional test pin can be difficult to implement and costly in terms of physical space, die size, and power requirements (such a test pin would consume unnecessary power because it would need to drive 50 Ohm test equipment).
A common method of testing the TIA bandwidth for an integrated TIA/limiter performs a low-level frequency response test on the TIA/limiter device. First, the test signal input level is reduced until the limiter is operating in the linear region. Then, conventional frequency response measurements are obtained with a spectrum analyzer and tracking generator combination, S-parameter test equipment, or a signal generator combined with a spectrum analyzer. Thus, the TIA bandwidth is measured in cascade with another linear system (the limiter). The frequency response of the limiter causes the measured bandwidth to be lower than the actual bandwidth of the TIA alone. The measured result is inaccurate because the limiter circuit normally operates in the nonlinear limiting mode, which does not significantly degrade the TIA bandwidth. In addition, the reduction of the test signal level results in a lower signal-to-noise ratio and noisy output measurements.
Another known technique for testing the bandwidth of a TIA implemented in an integrated TIA/limiter is set forth in a draft IEEE standard relating to Ethernet network systems (IEEE Standard 802.3-2002). This IEEE standard describes a test for an optical receiver module that includes a photodiode, a TIA, and a limiter. In summary, the test method is performed according to the following generalized steps: (1) sum a tone from an RF signal generator with a suitable data source; (2) with the RF tone set at the midband, adjust the powers of the RF tone and the data signal to provide a suitable bit error rate (BER); (3) change the RF tone frequency and vary the RF tone power to provide the same fixed BER; and (4) repeat the above steps for a number of frequencies to obtain the frequency response. The rolloff in the TIA frequency response corresponds to the increase in RF tone power as the frequency increases. Unfortunately, the IEEE testing technique is time consuming because several BER measurements must be taken to find the proper RPF power setting that yields the fixed BER at the given test frequency. If very accurate test data is desired, then the number of individual BER measurements at each frequency must be increased. In addition, BER test equipment can be expensive and difficult to operate.
Accordingly, a need exists for an accurate and effective technique to measure the TIA bandwidth in an integrated TIA/limiter device.
A testing technique according to the present invention can be utilized to test the frequency response of a TIA packaged with an integrated limiter, even though the integrated device lacks a TIA output pin. The testing technique can drive the TIA/limiter device at normal operating levels where the limiter functions in the nonlinear region. In this regard, the frequency response measurement is not affected by the operation of the limiter.
The above and other aspects of the present invention may be carried out in one form by a method of testing a transimpedance amplifier packaged with an integrated TIA/limiter. The method involves driving the TIA/limiter with a noisy input signal, obtaining a measured output jitter response of the TIA/limiter as a function of frequency of the noisy input signal, and deriving a frequency response of the TIA/limiter from the measured output jitter response.